Self-Routed Packet Network Architecture for Neural Computing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current packet networks face inefficiencies and unreliability due to slow routing processes, congestion, node failures, loop formation, and insecurity, especially in dynamic and large-scale networks, which hinder real-time decision-making and load balancing, and are not suitable for neural networks.

Innovation Solution

A Self-Routed Packet (SRP) network architecture that uses distributed hardware to enable fast, secure, and optimized path selection through hunting packets, token-based control, bridging, and pipelined processing, allowing for real-time, dynamic, and secure connection establishment and teardown of Label Switched Paths (LSPs) based on current network conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional IP routing is used in packet networks, then end-to-end packet delivery is supported, but routing processes are slow and cannot make real-time decisions

Engineering Contradiction:

Improverouting decision speedVSAvoidrouting process time

Core Design Contradiction:

SpeedVSLoss of time

Solution Approach 1:

The patent segments the routing decision process into distributed hardware-based routing tables at each node, rather than using centralized software-based routing protocols. Each node maintains local routing information in hardware, enabling parallel processing and wire-speed routing decisions without centralized coordination delays.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent replaces software-based routing protocols with hardware-based routing tables and switching fabric. This mechanical/electrical substitution enables deterministic, real-time routing decisions at wire speed, eliminating the processing delays inherent in software-based IP routing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If label-switching technologies like MPLS are used to improve packet network efficiency, then deterministic packet transport is achieved, but the network lacks intelligence for dynamic path selection and load balancing

Engineering Contradiction:

Improvepacket transport reliabilityVSAvoiddynamic path selection capability

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements self-service through distributed intelligence where each node autonomously makes routing decisions based on local hardware routing tables. The network automatically adapts to changing conditions through hardware-based path computation and dynamic rerouting capabilities, eliminating the need for external control planes.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent introduces dynamics through hardware-based routing tables that can be rapidly reconfigured in real-time. The switching fabric and routing logic can dynamically adjust paths based on network conditions, providing both the reliability of label-switching and the adaptability of intelligent routing.

Inventive Principle:

Principle #15Dynamics

3Adaptability or versatility

If software simulation is used to build neural networks, then neural network functionality is achieved, but the host computers are serial devices and therefore slow compared with true parallel neural networks

Engineering Contradiction:

Improveneural network functionalityVSAvoidprocessing speed

Core Design Contradiction:

Adaptability or versatilityVSSpeed

Solution Approach 1:

The patent replaces software-based neural network simulation with hardware-based implementation using the packet network infrastructure. The distributed switching fabric and hardware routing tables provide parallel processing capability, transforming the serial host computer model into a true parallel hardware system that achieves both neural network functionality and high speed.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

4Speed

If distributed hardware is used to enable fast routing decisions, then wire-speed path establishment is achieved, but device complexity increases

Engineering Contradiction:

Improvepath establishment speedVSAvoidhardware complexity

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

The patent achieves multi-functionality by using the same distributed hardware switching fabric and routing table structure for multiple purposes: traditional packet switching, neural network operations, and dynamic path selection. This universal hardware platform provides wire-speed performance across different applications without proportionally increasing complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10462039B2Data neural network system and method

Publication Date: 2019.10.29 WOOD SAMUEL FREDERICK

AI summary

This application discloses a neural network that also functions as a packet data network using an MPLS-type label switching technology. The neural network uses its intelligence to build and manage label switched paths (LSPs) to transport user packets and solve complex mathematical problems. This architecture is well suited to interconnect large numbers of processors or computers into a neural network exhibiting advanced intelligence which can be used for complex activities such as managing the power grid. However, the methods taught here can be applied to other data networks including ad-hoc, mobile, Information Centric, Content Centric, Sensor, and traditional IP packet networks, cell or frame-switched networks, time-slot networks and the like.